Studies on Plastron Respiration: I. The Biology of Aphelocheirus [Hemiptera, Aphelocheiridae (Naucoridae)] and the Mechanism of Plastron Retention

The probability that certain aquatic insects are able to extract oxygen from solution in the water by means of the air bubbles which adhere to their bodies was first suggested by J. H. Comstock in 1887, but it was not until Ege studied the subject in 1918 that the process was at all understood. Previously to Ege’s experiments opinion had wavered between the theory that either the air stores carried by aquatic insects were hydrostatic in function (Brocher), or that they were genuine stores of air on which the insect could draw for its oxygen requirements whilst submerged, most workers rightly holding a combination of the two. The work of Ege confirmed the suggestions of Comstock and of a number of subsequent observers that the air bubbles carried by such insects as Dytiscidae, Notonectidae and Corixidae, besides constituting an air store, can, in fact, be used as a kind of ‘physical gill Thus in many beetles a certain part of the surface of the subelytral air store is exposed to the water by abdominal movement. In other beetles, such as the Hydrophilidae and in the corixid bugs, an air film is carried entangled in the body-surface hairs, this film connecting with the spiracles and so providing a means whereby oxygen from the water can diffuse into the tracheal system. Insects which have such an air film entangled in their pubescence of course appear silvery when submerged. This system, consisting of the hairs and the layer of entangled air, was termed by Brocher (1912 b) the ‘Plastron’, a convenient term equivalent to Lufthulle of Schaufuss (1916). Gas-bubble respiration and plastron respiration in the original sense thus involve simply the transport of oxygen from the surrounding medium to the internal gaseous environment across an air-water interface formed by a thin film or bubble of gas held on the body surface by suitable hairs. Ege showed that some larger species of aquatic insects can only get enough oxygen by diffusion into their gas films or bubbles when they are quiescent, and that any considerable degree of activity involves frequent visits to the surface in order to renew the oxygen supply by means of pumping movements ventilating mechanically the air spaces and tracheal system. But in other smaller bubble-carrying insects, for instance, Hyphydrus and the smaller species of Corixa, the gill effect of their gas bubble or film supplies them with sufficient oxygen to render them practically independent of visits to the surface even when active. In the quiescence of winter conditions gas-bubble respiration may be sufficient to supply all needs, even of relatively large insects, and even under the ice there are always bubbles available for occasional renewal. The great value of Ege’s work, however, was that he showed clearly that insects which respire in this manner are subject to a serious danger if they stay below for too long a period without having access to the atmospheric air or to bubbles—even though the medium is saturated with oxygen at atmospheric partial pressure. He showed that the successful functioning of an air store as a gill depends on the fact that the ‘invasion coefficient ‘of oxygen between water and air is more than three times as great as that of nitrogen.^* Therefore, if the tension of oxygen in the water is higher than the partial pressure of oxygen in the air film there will be a tendency for equilibrium to be restored both by diffusion of oxygen into the bubble and by the diffusion of nitrogen out of it. This mechanism is probably of some value in almost all aquatic insects which have any exposed bubble surface, for Ege showed that a Corixa or Notonecta will, because of the oxygen invasion, benefit to the same extent as it would were the air volume which it carries about thirteen times what it actually is. In an oft-quoted experiment Ege showed that a Notonecta which lived 7 hr. in a quantity of water saturated with atmospheric air (access to the surface being denied) was only able to live 35 min. in water saturated with oxygen (provided that the animal was placed in an oxygen atmosphere for a while before the experiment began).

The accepted interpretation of this experiment is as follows. When the animal is in a large volume of water saturated with air, then the bubble will contain nitrogen as well as oxygen. As the oxygen in the bubble is used up by the animal the tension of oxygen in the bubble falls and the tension of nitrogen rises; oxygen therefore diffuses into the bubble and nitrogen diffuses out. In this way the animal is able to make use of the oxygen dissolved in the water. If the quantity of nitrogen in the bubble was constant this process could go on indefinitely, but since nitrogen diffuses out into the water the result is that the bubble becomes smaller and eventually water will enter the tracheal system. When the bubble contains pure oxygen and the animal is in water saturated with oxygen, there is no tension difference to cause oxygen to diffuse into the bubble. Under these circumstances the oxygen dissolved in the water is not available to the animal, which is therefore only able to survive until the oxygen originally present in the bubble is used up. In fact, in these smaller bubble-carrying insects in which this ‘physical gill ‘principle plays a major part, it may well be that the periodical visits to the surface are required rather to renew the store of nitrogen than to obtain a fresh supply of oxygen.

There are certain insects, however, which appear to have surmounted this danger and are able to hold gas in such a manner that it cannot under any normal circumstances dissolve in the medium and be replaced by water. It seems advisable henceforth to restrict the term plastron to the extremely thin gas layer held by these insects in contradistinction to the easily replaceable air films and bubbles (‘air stores’) of insects such as Hydrophilus

This arrangement differs fundamentally from an air store in that the volume of gas in the plastron is negligible and, moreover, the interface is held in position by surface forces so that the volume of the plastron is sensibly constant. The plastron therefore cannot function as an air store; its volume does not decrease nor does its air require to be replaced at the surface at intervals as the gas is used up. Owing to the tenacity with which the gas film is held in such an organ, differences of pressure may exist between the plastron space and the surrounding medium; therefore functionally it resembles more nearly the closed tracheal gill than the air store—even though the latter may to a limited extent also act as a physical gill.

The whole subject of plastron respiration in this strict sense is one of considerable theoretical interest and hitherto by no means understood. There are, briefly speaking, three groups of insects which show this phenomenon, but for various reasons they have hitherto all proved difficult to work with. Aphelocheirus (Hemiptera—Heter-optera, Aphelocheiridae) has always been described as extremely rare in western Europe. The same is true of the chrysomelid (Donaciine) beetles of the genus Haemonia {Macroplea), which have the additional disadvantage, for reasons which will appear later, of being extremely sluggish. The remaining group, Elmidae (Coleóptera Diversicornia), has many species which are common enough, but their very small size and again their sluggish habits render them very unsatisfactory for most types of experiment.

Brocher (1912b) showed that Elmis aenaeus was able to survive more than 50 days’ submergence even without the presence of gas bubbles in the water, provided, of course, that the water contained sufficient oxygen; and some desultory observations by other workers suggest the same thing in the other two groups mentioned. As long as the oxygen tension in the water remains sufficiently high these insects then appear able to supply their needs from the oxygen diffusion via the gaseous plastron. This raises a number of interesting theoretical problems—How is the plastron held? By what means is the danger of losing nitrogen avoided? As pointed out above the minute size of the Elmidae renders them difficult to work with, and because of their sluggish habits and their tolerance of considerable periods of oxygen want, they show no immediate and characteristic attitude denoting asphyxia. Haemonia, although larger, has the same objections and appears, moreover, to be excessively rare. Aphelocheirus is large and active and is therefore in many ways the ideal animal for experiments.

The senior author’s interest in Aphelocheirus was first aroused by Szabo-Patay’s (1924) account of the remarkable rosette-like structures, evidently the spiracles, which are found on the abdominal sterna of this insect. There are many similarities in respiratory adaptations between aquatic and parasitic insects, and Szabo-Patay’s quite gratuitous and, indeed, untenable suggestion that these rosette-like structures were concerned with selective removal of oxygen from the water recalled other physical assumptions implied in Maple’s (1937) discussion of the so-called ‘aeroscopic plate’ on the egg stalk of the parasitic Chalcidae (cf. Thorpe, 1936). It appeared probable from Szabo-Patay’s account as well as from the observations of some earlier workers that Aphelocheirus was an example of plastron respiration. Szabo-Patay saw that there was a fine hair pile over a considerable part of the body—such as might be expected to entangle a film of air for respiratory purposes (although this possibility did not apparently occur to him)—and it seemed probable that the rosette structures were in some way adaptations to this special mode of respiration. In effect, then, the opportunity to work on the subject of plastron respiration depended on the finding of a fairly copious supply of this apparently very rare hemipterous insect Aphelocheirus. A few occurrences extending over 70 years or more were on record for, in all, about ten localities in Great Britain and Ireland, but it was not until 1942 that Daniels and Ellis found Aphelocheirus abundantly in the River Wen-sum, near Norwich. On hearing of this record in 1943 one of us (W.H.T.) visited the locality and obtained some through the kindness of Mr E. A. Ellis of the Norwich Museum, to whom we have been indebted in many ways during the search for material for observation and experiment. About the same time Lt. E. S. Brown supplied valuable information about the occurrences of the insect in the River Cherwell, north of Oxford. For experimental purposes most of our material has been collected from one or other of these localities, the method being to don waders and walk slowly upstream through the shallow swift-running parts of the river, dragging a moderate-sized dredge, and disturbing the stones with one’s feet. The animals can be transported to the laboratory in tins containing damp waterweed, under which conditions they will, if required, live satisfactorily for as long as 2 or 3 days until they can be transferred to a properly aerated aquarium.

The present paper is the first of a series of four or five. In those which follow it is proposed to deal first with other aspects of the biology of Aphelocheirus in relation to respiration and to describe a detailed study of the respiratory efficiency of this insect and the theoretical principles involved in plastron respiration. This is to be followed by an investigation into plastron respiration in the Coleóptera and a study of the structure and function of insect hair piles in general, with special reference to the evolution of gas-bubble and plastron respiration.

Aphelocheirus is a large, tough, strongly flattened water bug, in appearance very unlike any other British aquatic insect. The form which occurs in this country is flightless, the hind wings being reduced to the merest vestiges hidden beneath the immovable semicircular forewings. A winged form, the status of which is obscure, is known from the Continent. A. aestivalis is normally characteristic of rivers or stretches of river with rapid current, well-oxygenated water, and a bottom of stones and small boulders on gravelly sand. It is a rapid swimmer, particularly at dusk, and if adults are kept in aquaria and observed at night they will be found to be highly active between 10 and 11.30 p.m., and occasionally later in the night or early in the morning before dawn. At such times they are found to swim close to the surface or clamber actively among the weeds. This activity is not the result of bad aeration of the aquarium water due to cessation of photosynthesis by the aquatic plants, but occurs equally in aquarium tanks supplied with a bubbler. During the daytime the bugs spend a large part of the time under stones, often just submerged in the sand. They do not appear to be associated with any particular kind of plant, although probably the presence of Potamogeton pectinatus is as good an indication as any of a suitable environment for them. They are carnivorous in habit and feed by probing with the long proboscis under stones and among crevices of rocks. In this way they obtain chironomid larvae and the larvae of caddis flies, e.g. Hydropsyche (Sirotinina, 1921) and the molluscs Vivipara vivipara (E. S. Brown in litt.), Cyclas and Pisidium (Larsen, 1924). Although Aphelocheirus is normally found in the type of environment indicated, it has also been recorded from a surprising variety of situations; Decksbach (1921), according to whom it occurs in many places in the Upper Volga, found it in shallow and stagnant puddles left by the subsidence of the main stream and states that in these collections of water the animal will persist for many months. Larsen (1932) has similar records. Beling (1926) describes it from many types of habitat in the River Bug in Poland, while Kulmatycki & Gabanski (1928) record that in the River Wierzyca larvae are able to resist successfully winter pollution caused by a sugar-beet factory, a pollution so great that it brought about a dominant fungus flora in the water. There are also records from brackish water near Kiel, as well as in Finland, Denmark and in France near the mouth of the Seine. It is known also from lakes, there being records from the English Lake District and from Lough Neigh, Ireland. A striking feature of the insect is the fact that while most frequently found in fairly shallow water, it has been recorded as going to a depth of 4 m. at Gudenaa in Denmark (Ussing, 1910), 5 m. in the Seine (Royer, 1913), and as much as 7 m. in the Volga (Decksbach, 1921, 1923). The eggs are laid mainly on stones and the life history is a long one, probably at least a year. All stages may be found at all times of the year. In Europe the species is apparently abundant in the Volga and some other Russian rivers, and is known from Sweden, Finland, Denmark, Poland, Germany, Austria, Hungary and France, but the general impression given by the records is that the insect gets progressively rarer towards its western limits. Szabo-Patay’s paper, for instance, was written as a result of seeing a single live specimen taken from the Danube. The distribution of the animal in the British Isles is somewhat remarkable and is not fully understood. One would think that such a relatively large (1 cm. long) and unusual looking insect—it suggests a huge aquatic bed-bug—could not fail to be recorded in most places where it occurred; but in actual fact the recorded localities even now (Bedwell, 1945) number less than twenty widely scattered areas in England, from Sussex, Hants, Berkshire, Norwich and Oxfordshire in the south to Berwick-on-Tweed in the north, Shropshire and north Wales (Pearce, 1945) in the west, and one in Ireland.

In appearance the insect is of a dark mottled brownish grey colour above, and underneath it shows a delicate pale grey sheen. There is no visible air store. There is, of course, no proper subelytral air space, but a little air is found under the rudimentary forewings. Since there is no bulk of air or air bubble carried the insect is heavier than water, and so, unlike many other aquatic insects, does not immediately bob up to the surface when it ceases swimming, but sinks to the bottom. Ussing (1910) pointed out that Aphelocheirus never normally comes to break the surface though it may swim about just underneath. It does not show respiratory movement of the abdomen or thorax.

Aphelocheirus will live and breed in aquaria, provided they are well aerated with a bubbler. We have also kept them in a special running water aquarium, but it is doubtful whether this has any real advantages over the simpler type. The animals will even live well for a considerable time in shallow dishes without any special means of aeration. As will be anticipated from the above-mentioned record of their survival of heavy pollution, they will survive for a long time in water containing much organic matter, provided they are not short of oxygen. If, however, the oxygen content falls the insects tend to swim more vigorously near the surface, often upside down, and may under such conditions try to climb out. If that fails they become sluggish and fall to the bottom, the hind end of the body then being raised slightly and the hind legs stretched out in a characteristic manner; these are the first signs of asphyxia. If oxygen want persists they then topple over on to their backs, giving spasmodic kicks—the second stage of asphyxia. Finally, they lie quiescent on their backs with their hind legs extended and crossed—third stage of asphyxia. If not left too long without aeration they will recover rapidly when transferred to aerated water. The nymphs are similar in general appearance but they show no ventral sheen, the skin is thinner and the whole insect somewhat transparent. They are more sluggish than the adult, but are capable of vigorous swimming. For a full description of the nymphal stages see Larsen (1927).

Szabo-Patay, in his 1924 paper, gives a good account of the tracheal system of adult female Aphelocheirus, and Text-fig. 1 is largely based on his paper, but has been corrected on some points in accordance with Larsen’s description (1924) and our own observations. Text-fig. 2 shows the tracheal system of the male in much greater detail. As will be seen from the figures, the tracheal system consists of two main tracheal trunks extending almost the whole length of the body. There are nine pairs of spiracles in all, two thoracic and seven abdominal, those belonging to the eighth abdominal segment having disappeared as in some other families of the Hemiptera (e.g. Cimicidae, Aphidae; see Weber, 1930, p. 272). The spiracles are of two entirely different types. The two pairs of thoracic spiracles and the first abdominal spiracles open into deep grooves running roughly along the sides of the much-flattened thoracic segments. The anterior thoracic spiracle is situated ventrolaterally at the junction between the pro- and mesothorax. From it a groove runs outward and upward meeting a small channel which connects with the second thoracic spiracle near the point of the attachment of the much-reduced mesothoracic wing, thereby connecting with the small air space between that wing and the dorsal surface of the metathorax and the second abdominal segment (i.e. the first abdominal segment to be well developed). The groove then runs back along behind the margin of the meta-thoracic subcoxal plate to meet the first abdominal spiracle at the point indicated. The groove becomes shallower and quickly fades out posterior to this point. These three spiracles each have a single very minute opening, and the groove with which they connect is closely packed with a thick-set pile of minute hairs. The structure of these spiracles has been described by Larsen (1924), but his account appears incorrect in certain respects. As will be seen from Text-fig. 6, the mouth of the trachea is protected by a screen of long cuticular filaments themselves beset with fine hairs. This mass of hairs may appear in section, under certain conditions, as a thick plate perforated with holes, and this no doubt explains the discrepancy between Larsen’s paper and our own observations. No closing apparatus appears to be present. The six abdominal spiracles situated on segments 2-7 inclusive are of an entirely different structure. They show ventrally as pale pinkish circular scars or patches which are made up of a number of fine branches radiating from the centre, the whole giving the appearance of a rosette, which name Szabo-Patay applied to them (‘Corpuscule en forme de rosette’). The appearance of a typical abdominal spiracle is shown in Text-fig. 3. The trachea which connects up with the main trunk can be seen underneath at about the middle point of the rosette. Sections reveal that each of the branching arms consists of a fine channel running through the substance of the exocuticle. As they pass from the centre outwards these channels become more superficial and open to the exterior by a number of very minute pores (see Text-figs. 3-5). There is a central scar near where the trachea is attached, but no opening is visible at this point. It will be seen from Text-fig. 4 that each of these grooves or channels is tightly filled with a hair pile which is, in fact, a continuation of the general hair pile which makes up the plastron of the ventral surface. At first inspection the spiracles give the impression of being closed, and it is not easy to make sure whether the tiny clear areas on the branches are in fact openings. That they are open is, however, clearly demonstrated by placing on the animal a minute drop of oil of cloves which has been heavily stained with Sudan black B. If watched under the binocular microscope the oil can be seen spreading along the surface and gradually filling the rosettes, and with a little care it is easy to see that the rosette branches fill from the tips and not from the centre. If the spiracle is now dissected out and mounted the drops of stained oil can easily be seen in the rosette channels. By this method the tracheal system can be completely injected with oil. As would be expected, the oil spreads more slowly down the arms of the rosette than it does across the ventral plastron hairs of the surface. Szabo-Patay describes an elaborate arrangement at the centre of the rosette consisting of a movable plate which he assumes is capable of performing a kind of pumping movement. On this he advances a remarkable theory of the function of the rosette which is, in effect, that oxygen is adsorbed on to the very great surface provided by the innumerable hairs, and that it is then removed as a result of the pumping action of the structure referred to and so rendered available for diffusion into the insect’s tracheal system ! We have been unable to confirm the existence of this pump either in stained sections or in whole unstained preparations. Study with the polarizing microscope fails to reveal any muscles which could be operating any pump. From this and other considerations which will appear later, we conclude that there is nothing in the structure of the abdominal spiracles to support Szabo-Patay’s theory of the mode of respiration. All the peculiarities of spiracle structure seem easily explained on quite another basis, and, indeed, Szabo-Patay’s hypothesis appears mechanically, physically and physiologically unreasonable.

It will be seen that the trachea running to the rosette of the second abdominal segment is somewhat thinner than those supplying the others, and that it discharges into a curiously shaped air sac the walls of which are folded and ribbed in a peculiar concertina-like manner. Its function will be discussed in the third paper of the series. The air sac opens directly into the rosette of the second abdominal segment, and that rosette is cut short by the presence on its outer edge of an oval plate, the lateral branches of the rosette running straight into it. This plate shows in the living insect as a brilliantly shining silvery patch. A moderately high power of the microscope reveals it to be a depression in the cuticle covered with long backwardly projecting hairs which entangle a relatively thick layer of air. It is, in fact, an organ of pressure sense of a remarkable type, the function of which will be discussed in a later paper. A conspicuous nerve runs to it from the composite ganglion which is situated in the region of the metathorax. The chief remaining features of the tracheal system can be seen from Text-figs. 1 and 2 and need not be described in detail. The only point worthy perhaps of note is the well-developed ventral commissures which connect the two rosettes of each of the abdominal segments 3-7 inclusive.

Larsen (1924) was the first to show that hairs similar to those found in the spiracular branches, but much shorter, cover the whole ventral surface of the animal as well as a considerable part of the dorsal surface. Owing to the dark colour of the dorsal surface the hair pile there is very hard to see. Ventrally, however, it gives a curious grey iridescent glint or sheen to the whole surface, whether the animal be examined in air or in water. Older individuals tend to get the hair pile scratched and worn and consequently appear darker, or with irregular dark patches. When the animal is removed from the water and the surface dried, it is noteworthy that the surface is not again readily wettable with clean water. This again seems to depend somewhat on the age of the insect, and one with a rather dark ventral surface will wet more easily—at any rate in the regions of the intersegmental membrane. A newly emerged adult, however, is strikingly hydrophobic, and water placed on the surface falls off as if the insect were covered with wax. An insect which has been thoroughly dried has no little difficulty in entering the water again, and does not usually manage to do so quickly unless it can pull itself down by getting hold of a piece of waterweed or some other convenient object. That the sheen of the insect is due to the existence of the hair pile containing air is very easily demonstrated. If it is touched with a drop of absolute alcohol or a solution of a wetting agent such as butyl alcohol or cetyl pyridinium chloride, the surface immediately darkens. This is evidently due to the wetting of the minute hairs and the consequent expulsion of the contained air. If an insect thus treated is rinsed and immersed in water without being allowed to dry, it remains dark in colour. If, on the other hand, it is allowed to dry thoroughly in air, the sheen reappears and the surface resumes its former unwettable condition. The extreme permanence of the gas layer of the plastron is indeed a very remarkable feature of Aphelocheirus. The plastron hairs will in many cases retain their ga^r layer almost indefinitely after death. We have had specimens which were collet;d in Norfolk in 1942, put aside in a corked tube containing a little muddy and highly polluted water, which showed their sheen almost undiminished two years later !

In the following section we show experimentally that the plastron in Aphelocheirus can retain a gas layer under all normal conditions and that it does, in fact, serve a respiratory function.

In considering the respiration of Aphelocheirus it is, of course, first necessary to make certain that it does not in fact need to come to the surface. To this end experiments on prevention of surface access were carried out, the animals being confined in a small glass tank (Text-fig. 7,1.T.), measuring 4 × 4 × 8 m. They were provided with a suitable layer of gravel and stones at the bottom, but were without any aquatic plants. This tank was closed by a tightly fitting perforated zinc grating (Z.G.). The whole was then immersed in a larger glass tank containing water to the depth of 5 in. There was then in. of water above the grating of the inner tank. A glass tube connected with the compressed air supply (A.S.) was led in through a hole in the zinc cover of the inner tank, this tube being bent round on itself at its tip (J.), which was inserted into another somewhat larger piece of glass tubing (L.T.). The two tubes were then fastened firmly together, and their lower ends protected by a perforated zinc guard (Z.g.) so that the bugs were unable to get up the larger tube or come into actual contact with air bubbles. When the air supply is turned on, bubbles then pass up the larger tube and out to the surface, drawing with them a steady stream of water, thus keeping the water in both tanks circulating and well aerated. A number of freshly collected female bugs were kept in this apparatus at a temperature of 18° C. the controls being in similar jars under normal aquarium conditions. It was found that the insects were able to live in an apparently healthy condition for at least 12 days, the only difference between the experimental and control animals being that the former appear perhaps rather more restive than the latter. There was no change in the appearance of the plastron during the experiment. It can be safely assumed, of course, that if nitrogen loss had been occurring during this time the gaseous plastron would have completely disappeared. This would, in fact, be expected to occur in the first few minutes if not sooner, since the volume of gas in the plastron is so extremely minute. Brocher has described in Haemonia (1912a) that minute air bubbles extruded from a micropipette will adhere to the plastron and be readily absorbed. We have not been able to observe anything of this kind in Aphelocheirus. Bubbles will sometimes cling for a little while to the marginal setae or to the leg bristles, but there was no adhesion or absorption as described by Brocher.

The next step was to see whether the insect is in danger of losing its gas layer even when in gas-free or oxygen-saturated water. It will be remembered that Ege has shown that a Notonecta placed in an oxygen atmosphere and then moved to oxygen-saturated water but denied access to the surface very rapidly loses its bubble; for the oxygen is absorbed, the carbon dioxide diffuses out and immediately dissolves, and in the absence of nitrogen no tension difference can be established between the bubble and the medium. Experiments of this kind with Aphelocheirus were quite conclusive. If insects are placed in a litre jar of water which has first been boiled gently for an hour to remove all gas and then cooled and saturated with oxygen from a bubbler for a similar period, they remain perfectly normal, no change or abnormality in their behaviour being observable as compared with control insects. If the insects are placed in gas-free distilled water they respond immediately by great activity, swimming about at the surface trying to climb out of the tank, and sometimes succeeding. If prevented from climbing out they cling to the side, and if they accidentally drop down, immediately come to the surface again. After a few minutes their activity decreases and they fall to the bottom and soon pass through all the stages of asphyxia mentioned earlier. The final stage of asphyxia supervenes in about 35 − 40 min. at a temperature of 18 ° C., but bugs removed at any time up to about 2 hr. may show quite rapid recovery. More than about 2 hr. total deprivation of oxygen is sufficient to kill the animal, but during such an experiment there is no visible change in the plastron. To follow this matter further, the following experiments were performed.

A bug with a very good ventral sheen was decapitated, and the right half of the abdominal sterna of the last four segments was dissected and placed in 20 c.c. of freshly boiled distilled water under a ‘Nujol’^* seal. The left half of the same individual was placed in aerated distilled water under the same conditions. The experiment was continued for 144 hr. but no differences between the two were observable. A similar experiment was conducted with living insects. Freshly caught females with a very good ventral sheen were fastened on their backs on a cover-slide with a small quantity of Euparal. One specimen was placed in a conical flask containing 250 c.c. of freshly boiled distilled water run in and cooled under a ‘Nujol’ seal. Controls were kept in exactly similar conditions except that the water was well aerated. The experimental animal was thought to be dead in about 3 hr., but no change could be detected between the sheen of the experimental and of the animal controls even after 120 hr. That Euparal has no ill effect on the animal was shown by the fact that one of the controls after being released from its cover-slip was living quite normally in the aquarium more than 3 weeks later. The experiments were repeated several times under the most rigorous conditions, using gas-free ‘Nujol’ and also using water kept ‘boiling ‘under a filter pump at room temperature for the duration of the experiment. In these instances dead insects were of course used.

The next step is obviously to show whether or not the plastron is acting as a gill. For this purpose a simple procedure was evolved of rendering a part or the whole of the plastron ineffective by ‘painting out’ the plastron under the binocular microscope with a camel-hair brush wetted with cetyl pyridinium chloride or cetyl trimethyl ammonium bromide. This removes the gas layer completely where it is applied and does not show any tendency to spread, providing the insect is quickly rinsed with distilled water. Although these salts are somewhat toxic, a brief application, using a 0·5 −1 % solution, followed immediately by rinsing in rapidly running tap water, generally avoided pathological effects. Immediately after operation the bugs were seen to lose some limb co-ordination and to adopt the pose of acute asphyxia, but generally recovery was complete after about 1 hr. The animals were treated on the previous day and allowed to recover overnight, so that any which were damaged and subsequently died were not used in the experiment. Treatment with negative soaps such as sodium dodecyl sulphate and sodium cetyl sulphate was less effective, and concentrations high enough for painting out the air film were found to have definitely toxic effects. Insects thus treated with a positive soap were observed in normal conditions and also subjected to water of known oxygen content less than saturation at atmospheric partial pressure. It was first found that if the whole of the ventral surface of a female were wetted in this way and the animal transferred to 250 c.c. of well-aerated tap water, the insect would become quiescent and show occasional symptoms of asphyxia, but not of acute oxygen want. Controls in which the dorsal surface only was treated showed no significant difference in behaviour from normal animals. The methods were as follows :

A series of four small jars (capacity 200 c.c.), containing 100 c.c. distilled water and a bottom layer of fine sand, were connected with the compressed air system via a mixing chamber to which air and nitrogen were led at a standard rate and constant head through flow-meters. The flow-meters were so adjusted as to give the required gas mixture. Thus, in the experiment where the water contained only 3·5% of oxygen, the air was led through at 2·4 c.c. and the nitrogen at 12·1 c.c./min. This air flow was passed through each of the four containers for some hours before the beginning of the experiment so as to be sure that equilibrium had been established. In the first jar (A) were then placed four bugs which had had the whole of the abdominal and thoracic sterna treated, care, however, being taken to avoid penetration of the spiracles by the wetting agent. In the second jar (B) were placed the three bugs which had had the abdominal sterna treated, and in the third jar (C) three which had had the whole of the tergum only treated. In the fourth jar (D) were three control insects. The animals were then observed at intervals of a few minutes and the behaviour classified as follows: N, normal; LA, light asphyxia; AA, acute asphyxia; C, comatose; M, apparently moribund, i.e. completely relaxed and giving no responses. These and many other similar experiments are summarized in graphical form (Text-figs. 8–12) by giving arbitrary scores to the different degrees of asphyxia as follows: N = 20, LA= 15, AA= 10, C = 5, M = 0. (Since the curves in Text-figs. 8–12 each refer to three individuals, these numbers are multiplied by three.) It will be seen that in most experiments there is initially a period of depression of the treated bugs, which can perhaps be regarded as one of shock. After this is over, however, a clearly marked difference is observable between the experimental and the control animals. Those which have the abdominal and thoracic sterna treated are most seriously affected, those which have the abdominal sterna only treated are only slightly better off, while those which have the whole tergum treated show very little difference from the controls. Text-figs. 10 and 11 show results from a series of experiments where bugs were exposed to mixtures in which the oxygen contents were 5·4 and 3·4% respectively. At the conclusion of the experiments all the bugs showed rapid recovery on removal to well-aerated water. In general, those which were less drastically treated and slowest in showing asphyxia were the first to show recovery. Rates of recovery were, however, very erratic, since without constant stimulation it was not possible to determine at what stage the animal was again capable of movement or of assuming a normal attitude. In observing the results of the experiments it was noteworthy that the differences between the animals was often rather a degree of activity than an actual asphyxia attitude. Since it was not possible to express this adequately in the scores the results are in fact more striking than the curves based primarily on the attitude would indicate. From these experiments we can conclude that simple body-surface respiration, without the intervention of any plastron, will suffice, under good conditions, when the animal is quiescent, for the prevention of all except the mildest degree of asphyxia. Such respiration will include, of course, a small amount of invasion at the spiracles themselves, since these were not actually blocked. But an intact plastron becomes vitally important for the maintenance of normal activity and for survival when the oxygen tension is appreciably reduced below saturation at atmospheric partial pressure. These results were fully confirmed by respirometer experiments to be described in the next paper in the series.

Text-fig. 12 shows for comparison the tolerance to oxygen deficiency of well-grown fifth instar Aphelocheirus nymphs studied in the same way. As will be shown in a subsequent account of the tracheal system of the nymph, the respiration of the early stages is on a different basis from that of the adult. The former has a completely closed tracheal system and lacks the plastron, and, being smaller and very flat, exposes a large surface to the water. In addition, the lateral margins of the abdominal segments are greatly produced and richly supplied with fine tracheae, thus presumably acting as tracheal gills (see Text-figs. 13-17). The cuticle, too, is distinctly more transparent than in the adult. The difference in weight between a well-grown fifth instar nymph and an adult is, of course, considerable; thus an average weight of a living adult, air dried, is 0·04 g. (average of five individuals), whereas a well-grown fifth instar nymph, not yet ready for moult, weighs about half this amount. The nymphs are on the whole less active than the adults, and it appears that when the insect is of the size at which it is due to assume the adult condition, a critical period is reached, at which simple diffusion across the cuticle is no longer adequate for respiration. This question will be considered in greater detail in a later paper.

In order fully to understand the way in which the plastron hairs hold the gas, and in order to appreciate the limitations of plastron respiration, it is necessary to know the exact shape, number and dimensions of the plastron hairs. To determine this has proved to be a much more difficult matter than was at first expected. The hairs are not stainable but show in section as pale yellow structures. They are so minute as not to be clearly visible except under an oil-immersion lens with a high-power ocular. They show rather poorly in balsam, and somewhat better in Euparal and better still in biniodide of mercury and potassium, a medium which has been found valuable for examination of minute detail in protistan skeletons. The usual method of examination was to make sections with the freezing microtome and to examine them under an oil-immersion lens with an Ediswan 150 c.p. ‘Pointolite’ and a Wratten H Filter. Under these conditions a very close-set pile of hairs is visible (Text-fig. 18), arranged with the utmost regularity and neatness. (In many parts of the section the air will, if alcohol or any other wetting agent has not been used, still be present in a large part of the hair pile, and there, of course, the hairs cannot be seen.) Under the best conditions of mounting and lighting and with a high-power ocular the hairs can be counted and drawn, but even so it is impossible to arrive at a convincing measurement of their thickness, which appears to be rather less than the wave-length of ordinary light and therefore at the limit of resolution. It is, however, clear (see Text-fig. 19) that where unharmed they stand perfectly erect, making an angle of 90 ° with the cuticular surface, and that they are neatly bent over at the tips, all in the same direction. A series of careful counts of sections gives an estimate of their numbers at approximately million per sq.mm. In certain protected places, e.g. the spiracular grooves of the thorax and the channels of the spiracular rosettes (see Text-figs. 4, 20), the plastron hairs are longer than they are on the more exposed surfaces of the thoracic and abdominal sterna, but the essential structure appears the same. In order to get further information as to the thickness and dimension of the tips of the hairs, studies were made with the electron microscope. With this instrument two methods were employed. First, from among a series of sections freshly cut with the freezing microtome, some were picked out which tapered toward the edge and thus tended to have very thin margins. These were teased and pulled with fine needles under the dissecting microscope so as to increase the chance of separating out single hairs and the section was then placed on the grid. Here and there the tip of a single hair was visible under the electron microscope as a silhouette (Pl. 7, figs. g, h). Secondly, the surface of the abdominal sterna was studied by means of the gelatin-formvar replica technique. In using this method the sterna were covered by a thin layer of 10 % solution of gelatin and left to dry hard. The gelatin was then cracked off and coated very thinly with a solution of formvar in dioxan, which dried rapidly. The specimen was then placed in slightly alkaline water to dissolve the gelatin away, and the formvar film floated on to the grid for examination. A true replica was not obtained, since the hairs are extremely long and thin for such a method. On cracking off the gelatin, however, the embedded hair tips were plucked out, and a considerable number attached themselves to the formvar film. Pl. 7, figs, g and h, show the features which are revealed by these methods. In spite of the many difficulties encountered they show fairly clearly the tips of a number of hairs and the diameter works out at 0·21−0·25 µ. Because of the difficulty of obtaining sections thin enough to be transparent to the electron beam, and because of the difficulty of orientating material in the field of the electron microscope, the hair structure was also investigated by means of an ultra-violet light microscope.

This has given very satisfactory results and has confirmed the previous estimate of dimensions and number of the hairs. Pls. 6 and 7, figs, a-f give a selection of photo-graphs of the plastron hairs taken by this method. From this it is estimated that the diameter of the hairs on the ventral surface is 0·18 µ, the distance apart (i.e. the distance between cores) 0·6µ, and the space between adjacent hairs approximately 0 ·4 µ. Counts based on the ultra-violet light photographs give a result similar to that obtained with a visible-light microscope, namely, million per sq.mm., although some estimates based on a surface view under oil immersion suggest that in some regions the density is less than this, being round about 2 million. The total area of ventral surface of the adult, including the head, is approximately 36 sq.mm. The hairs on the dorsal surface are shown in Text-fig. 21. Here they are often not so regular and do not appear to bend over at the tips in the same neat way as those found on the sterna, though in some regions they appear slightly clubbed at the tips (see Text-fig. 22). The whole cuticle is much more deeply pigmented, with the result that the terga have not the same sheen as the sterna, but it is doubtful if there is any fundamental difference between the two.

It will be seen from Text-figs. 18 − 22 that in the adult there is a thick endocuticle lying above the hypodermis, this endocuticle being composed of five or six distinct layers, the outermost of which shows rather distantly spaced cross-striations. The endocuticle contains practically no pigment. The exocuticle is heavily pigmented. In some sections little structure can be seen in it, but in others innumerable fine transverse lines can be observed, which appear to be pore canals. These are shown in Text-fig. 19. Only rarely is it possible to see where these pore canals enter the endocuticle. The epicuticle is very difficult to distinguish from the exocuticle, but it appears in most sections as a faint line bounding the surface from which the plastron hairs arise. It is of course impossible, by ordinary microscopical methods, to observe any structure in the plastron hairs themselves, because of their minute size. Campbell’s concentrated KOH-H₂SO₄-KI method produces a good chitin reaction both in the exo- and endocuticle but none at all in the epicuticle or plastron hairs. The coloration in the exo- and endocuticle is quite general and does not appear to be confined to pore canals. In no case was staining of the pore canals obtained, yet in a few cases in sections of adults, pore canals were rendered visible in the endocuticle after treatment with 1 % H₂SO₄, following the potash treatment applied in Campbell’s method for chitin. The best of these sections were photographed and both exo- and endocuticle pore canals can be made to show well, although a different exposure is necessary for the two. It appears from these photographs that pore canals are more numerous in the exocuticle than in the endocuticle. It is difficult to be sure whether or not this means that the canals branch at the junction of the two layers. In many sections pore canals could be seen in some layers of endocuticle but not in others. Often they were conspicuous in alternate layers, but it is not at present clear what significance this may have. It is interesting to note that Dennell (1946) finds pore canals present only in the older outer layers of the endocuticle of the larva of the fly, Sarcophaga falculata, no trace being observable in the inner layers. The interpretation of sections after this drastic potash treatment is often difficult. The exocuticle It is interesting to note that endocuticle, and different layers of endocuticle tend to split and form a tangled mass of almost structureless fibres. Yet in spite of this the plastron hairs will often survive the treatment, clumped and appearing partly fused but still visible. Little is known then of the physical properties or of the chemical nature of the epicuticle, of which the hairs appear to be composed. From its. great resistance to chemical action it is likely to be a highly cross-linked substance, most probably protein, with perhaps some chitin.

As will be seen from some of the photographs taken by ultra-violet light there is often an appearance which suggests that the plastron hairs are essentially continuations of structures within the exocuticle. Pls. 6 and 7, figs. c, d, e, in particular, show what appear to be the ‘roots ‘of the plastron hairs passing through several horizontal layers in the outer part of the exocuticle—layers which are not visible at all by any other method we have tried. The fact that these ‘roots’ absorb ultra-violet radiation more strongly than their surroundings suggests a protein framework. It is interesting in this connexion that Hurst (1945) on the basis of permeability studies has postulated a mosaic structure of the insect cuticle very much of this form. It is also likely on general grounds that there is a molecular orientation along the axis of the hair. The structure of the nymphal cuticle and its relation to the question of development of the plastron in Aphelocheirus will be discussed in the second paper of this series.

It was noted quite early in experiments with Aphelocheirus that the tips of the plastron hairs of newly emerged adults, that is, the surface of the plastron itself, seemed more hydrophobic than with older animals. It was also observed that if an adult bug was taken from water and carefully dried with filter-paper and then kept dry for half an hour, the hydrofuge properties of the plastron surface seemed enhanced.

As mentioned above cetyl pyridinium chloride proved to be a satisfactory wetting agent when it was desired to remove the plastron or any part of the plastron as a respiratory structure without harming the animal. It was instructive for some purposes to watch the spread of oils such as clove oil and medicinal paraffin (Nujol) over the plastron surface, but these were not satisfactory for experiments in that their spread could not be conveniently arrested. A number of experiments were carried out to determine the wetting power of solutions of pure isobutyl alcohol in water at 18 ° C. It was found that solutions up to 8% strength had no effect whatever. Above this strength, changes in appearance of the plastron might commence slowly but had not proceeded far in 5 min. Solutions between 9 and 11% had a more marked effect, in some specimens the sheen disappearing fairly quickly at 10%, in others very slowly, but even in the most resistant specimens the change proceeded rapidly at 12% and was easily noticeable after 30 sec. exposure. Full wetting with butyl alcohol caused a dull black colour to replace the sheen, and no restoration could be achieved even in tap water (supersaturated with air). When washed, dried and reimmersed in water the sheen was fully recovered and no change in properties could be noticed, resistance to wetting being in no way diminished. Hence it is not likely that butyl alcohol exercises any irreversible change on the structure of the hair pile beyond normal wetting. The well-known difference between advancing and receding contact angles is sufficient to explain the irreversibility of the process as long as the system is kept immersed.

It has already been shown that the plastron hairs are not wetted by immersion in gas-free distilled water. The next question is, what hydrostatic pressure is necessary forcibly to wet the plastron with gas-free water? In order to determine this, a piece of abdomen including sterna of the anterior segments was subjected to increased pressure in distilled water in an apparatus consisting of a stout pressure jar and a manometer. In order to observe accurately the change in sheen, it was necessary to have the specimen appropriately orientated in good diffuse daylight (various sources of artificial light were tried but all were inferior to daylight) and to have a control specimen in a similar chamber beside it for comparison, orientated in the same way. Experiments were repeated many times, and it was found that with a fairly young bug with a good sheen the glint began to fade at about 3 atm. excess pressure, although the margins of the segments were still at full sheen. The sheen vanished entirely and immediately at a pressure of between 4 and 5 atm., but it was noticeable that the specimen did not acquire the dull black appearance which is produced by wetting with butyl-alcohol solutions. If the pressure was released and the specimen now transferred to aerated distilled water or to tap water, some recovery of sheen was shown in about 10 min. Complete recovery did not appear to take place as long as the specimen was immersed. If, however, it was dried and then returned to water, its appearance became completely normal again. In another experiment the specimens were kept at a pressure of 2 atm. for a longer period. In this case no marked reduction in sheen took place, though there was a barely perceptible change after a period of 5 hr.

When the hair pile is exposed to increased pressure in a fluid of lower contact angle and surface tension than pure water, wetting should take place at a lower value than those just described. The experimental points on Text-fig. 26 were obtained by exposing the bug to excess pressure in distilled water saturated with air and containing different concentrations of butyl alcohol, whose contact angles against wax had previously been determined (see Text-fig. 25).

The difference in appearance and in tendency to recover sheen as between plastron wetted by fluids of low-surface tension on the one hand and by pure water under excess pressure on the other, indicates that we have here two entirely different phenomena. In the first case we have a true wetting brought about by the entry of the fluid in between the hairs thus displacing the gas. In the second case the evidence strongly suggests that the excess pressure causes a collapse of the hair pile itself.

When external pressure is applied to the hair pile it is not only necessary that the surface forces should prevent entry of water into the plastron, but it is also necessary that the hair pile itself should be capable of withstanding the pressure difference. If this were not so the hairs would collapse and the air space be eliminated.

We shall now examine the conditions under which a system of hairs would be theoretically capable of maintaining an air film against excess pressure outside, and calculate the order of magnitude of the pressure difference involved. The reader is referred to N. K. Adams Physics and Chemistry of Surfaces, 3rd ed. 1941, pp. 177201 (Clarendon Press), for the general principles underlying this section.

The second system which will be considered is that in which the hairs are cylindrical, and lie in a parallel series as shown in Text-fig. 23 c, d.

It can be shown from the geometry of the system that, since the amount of surface of hair covered by the water varies non-linearly with the lowering of the water surface, the excess pressure in equilibrium with the surface forces gradually increases as the water surface is lowered and the water ‘bubble’ is forced into the space between the hairs. This excess pressure passes through a maximum and decreases again, just as the inflation through a tube of a spherical bubble passes through a pressure maximum.

βbeing the angle between the horizon and point of wetting of the hairs, and assumed positive when more than half the hair is wet (see Text-fig. 23 b), and θ the contact angle between hair surface and water surface.

The value of Δ and γ must next be considered in applying these equations to the hair structure of an insect.

Since the hairs have a hydrophobic surface, either on account of their innate structure or due to a waxy layer, the outer surface is likely to be composed mainly of hydrocarbon groups. The contact angle is dependent only on the molecular configuration of the outer layers, and if these are hydrocarbon the value of θ lies in the region of 105−110°. If the surface is rough or cavitated there will be an apparent increase in 0 (Wenzel, 1936; Cassie & Baxter, 1944), but this factor will be ignored for the present and a value of 110° taken, the surface being assumed to be similar to that of solid hydrocarbon. Electron microscope pictures of the hair indicate no roughness of their surface.

This value of θ refers to the contact angle against a clean water surface. If the surface tension of the water is altered, θ will change on account of the alteration in the forces of the Neumann triangle, but the change in θ is not in the expected sense. In Text-fig. 24 it can be seen that

where γ_s is the solid-air tension, where γ_LS is the liquid-solid tension, where γ_L is the liquid-air tension. A reduction in γ_L caused by an impurity would be expected to cause θ to increase if θ > 90°. In practice, however, θ always decreases, probably because adsorption occurs simultaneously at the liquid-solid boundary, decreasing γ_Ls. It will be observed from the graph (Text-fig. 25) that the adhesional work (γ_L+γ_s—γ_SL) varies slightly and linearly with the surface tension over a greater part of the curve, suggesting that adsorption is similar at the air-water and hydro-carbon-water interface. The relation between γ_L and θ is not necessarily similar for all impurities, but depends on the equilibrium values of γ_s and γ_LS when an impurity is present. The θ-γ relation for three insoluble oils in also given.

(Bikerman, 1940), where Δ = length of base of droplet and v= volume of droplet measured by an Aglar microsyringe.

The surface tension was measured by the ring method applying Harkins’s corrections (Harkins & Jordan, 1930). In order to apply the equations to conditions under which θ might vary on account of impurities we made the following assumptions :

• That the surface of the hair resembled that of a freshly cut block of pure paraffin wax of m.p. 56° C.

• That butyl alcohol may be taken as a typical surface active impurity.

Applying these values of θ and γ to equations (1), (2) and (3) we obtain the curves shown in Text-fig. 26, assuming the dimensions r = 1 × 10⁻⁵,1= 5 × 10⁻⁵ cm. for the hair pile of the bug Aphelocheirus.

It is apparent from the graph that the least efficient type of hair pile is one in which the hairs are directed vertically to the surface (curve 1). Curves 2 and 3 (a = 45° and 30°) indicate the improved efficiency against wetting which is obtained by inclining the hairs at an angle. The hairs are still wetted when θ is < 90°, however, and there is a limit to the angle α, since eventually the hairs would touch one another. For instance, if the r/l ratio is 1:5, the hairs would touch if inclined at an angle of less than 25°.

Curve 4 indicates the pressure necessary to wet a system of purely horizontal hairs; it will be noted that a much greater pressure is required to wet the hairs, and that a positive pressure is necessary for all contact angles above zero. This is clearly a far more efficient system.

In practice, of course, it is not possible to have a system of horizontal hairs without support. Vertical hairs in which the tips are sharply inclined at 90° would offer the best approach to such a condition and this is what is in fact found in Aphelocheirus.

The effect of the vertical supports to the horizontally placed hair tips will naturally cause the behaviour to show some deviation from that predicted from equations (3) and (4) and shown in Text-fig. 26, curve 4. We should anticipate that on increasing the angle θthe line of contact of the water surface will not remain straight when β> o, but will assume a contour roughly as shown in Text-fig. 23,d. It is impossible to apply any mathematical analysis to this system, but it is clear that for contact angles θgreater than 90°, the vertical portions will draw the water surface downwards and so cause instability and wetting at a value of β which is less than 90°, possibly much less. When this unstable point is reached the menisci will follow the vertical portions and cause immediate wetting and waterlogging of the plastron. If θ is greater than 90 °, however, the vertical regions will also resist wetting and their effect will not be to cause instability. Text-fig. 26, curves 5 and 6, show the effect if the system is assumed to become unstable at a value of β of less than 90 ° ; curve 5 is for β < 50 °, and curve 6 for β < 30 °. The experimental points shown on the same graph were obtained as described on p. 255 above. They indicate clearly that the hair pile functions as a horizontal system, but instability rapidly sets in as predicted above, when the contact angle is less than 90 °. Above 90 ° contact angle the points agree very well with those calculated from equation (3). Although the precise agreement found is probably somewhat fortuitous, since Δ p is very sensitive to the density of packing of the hairs (which will be altered if there is any considerable overlap), there is no doubt that agreement exists both in respect of the order of magnitude of Δ p and the anticipated relation between Δ p and θ.

When θ falls below 90° the agreement breaks down and wetting occurs at zero pressure when θ is just above 60°. This is doubtless due to the presence of the vertical supports. It will be seen that the experimental points in this region agree well with curve 6, calculated on the assumption that instability sets in for values of β > 30°. This limitation of β has relatively little influence on Δ p when θ is greater than 90°, so that curve 6 offers a good approximation to the observed behaviour throughout.

As has already been shown the resistance of the hair pile to collapse is a most important factor which must be taken into account in considering the mode of action of the plastron. It is possible to calculate the order of magnitude of the forces involved when external pressure is applied to the hair pile if a reasonable value for Y (Young’s modulus) of the epicuticular hairs can be assumed. From the evidence discussed above we think it justifiable to assume that the hairs include in their make-up a highly cross-linked protein framework with perhaps a molecular orientation along the axis.

Values of Y for a series of substances of polymeric character likely to be of the same order are shown in in Table 1.

This table indicates that, while amorphous linear plastics have a low Young’s modulus of about 0· 2− 0· 4 ×10¹¹ dynes/sq.cm., cross-linked plastics containing the CO.NH bond are much stronger and have values of the order of 1·0×10¹¹. Wood, which contains orientated crystallites of cellulose and some cross-linking, is of the same order; if allowance is made for the voids, the strength would be about 50% greater. These figures suggest that for a well cross-linked protein framework the modulus would be in the order of 0· 5−1·0×10¹¹ dynes/sq.cm., and possibly a little higher if the chains are orientated.

We shall use this value as an approximate order of magnitude. The length of the hairs is h = 5 × 10⁻⁴ cm., and the radius =1· 0×10⁻⁵ cm. The latter figure introduces the largest error since it is raised to the fourth power. In ail the insects studied where the hairs are easily visible (e.g. Haemonia, sense organ in Aphelocheirus, Elmis) there is a slight widening at the base of the hair giving greater rigidity; hence r= 1· 0 will be a minimum value.

To calculate resistance to compression two methods were employed:

• Euler’s treatment. The hair is regarded as a flexible rod subjected to compressional force causing buckling. When the applied force is critical the rod bends to a stable position so that

  

  where F=critical force, Y=Young’s modulus, A = area of cross-section, K=radius of gyration of cross-section about neutral filament, and h = length.
  For a cylindrical rod⁠, and for F we may put Δ P/n, assuming equal distribution of the pressure over n hairs/sq.cm. (n=2—3 × 10⁸). Therefore

  
• Treat the hair as a rod with distributed weight, the weight being

  
  Calculation indicates the critical condition of stability is where

  
  (Champion & Davey, 1936). Hence

  

  There is reasonable agreement between the two calculations.

These values indicate that the critical applied pressure at which the hairs will bend is probably about 2·5−6 atm. This is very close to that at which wetting occurs in clean water. However, as the hairs are bent down under pressure, they will eventually lie over one another, since the area of the hair lying flat is 5 × 0·2µ²=1µ², which is greater than the area allowed per hair, 0·3−0·5µ². In this condition they will not only be able to exert a much greater mechanical resistance to compression, but their proximity will increase the resistance to wetting (vide pp. 261, 264−5). But owing to the squeezing out of most of the air, the plastron will no longer be able to function as a gill and may not show the typical sheen.

Evidence that bending of the hairs did actually occur was obtained during pressure experiments. When pure distilled water was employed, the sheen did not darken completely, but ‘flashed’ darker on the initial stroke of the pump at about 2·5 atm. excess pressure. At 3·7 atm. it remained dark all over but did not acquire the dull black colour of specimens wetted with butyl alcohol. Moreover, immediately on release of pressure the sheen partially returned. With 0·5% butyl alcohol (θ= 106−107°) the phenomenon of ‘flashing’ just preceded wetting and the return of sheen was only just visible.

It is also interesting to note that when darkening occurred on account of the hairs bending over, this took place first in the mid-line of the sterna where the hairs are longest, and only at a higher pressure at the lateral borders. Such would be expected, since it will be seen from equations (5) and (7) above that Δ p is proportional to h^-2, where h is the length of the individual hairs.

The pair of sense organs which lie on the outer posterior border of the second abdominal (apparent first abdominal) sternum admirably illustrate the above-mentioned waterproofing properties of closely packed adpressed hairs. Their size is about ten times that of the body hairs, they are very long (Text-fig. 27), and lie parallel to one another so that they overlap considerably.

Where this hair pile is in contact with the water surface the hairs, if free to do so, will tend to concentrate at the interface (Text-fig. 27 a, b), since in this way the area of free water surface is reduced. In the limit the hairs will be touching each other at their diameters (Text-fig. 27 c); this will take place if a downward pressure is applied to the interface to bring a sufficient number of hairs into it. In this system the equations developed for a horizontal array of hairs may be employed (equations (3) and (4)), but instead of a fixed distance I between adjacent hairs, the distance I will be gradually reduced to a limiting value of 2r as the hairs become concentrated at the interface. In Text-fig. 28 the maximum excess pressure Δp above such a system is plotted as a function of the distance separating the hairs. &p is expressed in atmospheres and the dimension of the hairs is taken as 1 × 10⁻⁴ cm. radius as found in the sense organ. The values of γ are related to θ as in Text-figs. 24 and 25, being based on measurements of contact angles of butyl alcohol on wax blocks. It will be seen that for contact angles greater than 90° there is no limit to the value of Δp, provided the hairs are sufficiently close; for smaller contact angles Δp has a maximum value when the hairs are in contact, which is of the order of 1 to a small fraction of an atmosphere, but always positive. As will appear from a succeeding paper, although the hairs are much larger and on that account present relatively smaller surface forces, the close packing will account for their greater resistance to wetting. This resistance should be reduced sharply when θ falls below 90°. If the hairs were of the same radius as the body hairs the magnitude of Δp would be ten times larger; but, although well waterproofed, the air-water surface would be so reduced that the structure would not function as a respiratory organ,

It is a pleasure to acknowledge our indebtedness to Mr E. A. Ellis, Naturalist at the Norwich Museum, for generous assistance and advice in discovering localities for Aphelocheirus and Haemonia. Similarly, we are indebted to Lt. E. S. Brown for much valuable information as to the behaviour and habitat of Aphelocheirus in the River Cherwell. To the authorities and staff of the Mount Vernon (Hampstead) Laboratory of the Medical Research Council and of the Cavendish Laboratory, Cambridge, we are indebted for unstinted help and facilities in the use respectively of the ultra-violet photomicrograph and the electron microscope. Finally, we are greatly indebted to Prof. E. K. Rideal, F.R.S., for first suggesting that D.J.C. should collaborate in this investigation.

PLATE 6

Fig. a. Sections showing plastron of abdominal sternum of adult Aphelocheirus. Sections cut with freezing microtome, mounted unstained in glycerine and photographed with ultra-violet radiation. Aperture 5 mm., exposure 5 sec. Note that with this exposure only the hair pile is visible, the epicuticle and exocuticle being impervious to the radiation and appearing completely black.

Figs, b, c and d. Same section as a with much greater exposure to show structure of epicuticle and exocuticle.

b.Note that the plastron hair pile is now so over-exposed as to be invisible, but the ‘roots ‘of the hairs can now be seen passing through five of six separate lamellae which appear to constitute the epicuticle. The exocuticle is still almost completely opaque. Exposure 30 sec., aperture 5 mm.

c.Exposure 1 min. Note that the exocuticle is now partly transparent and that the ‘roots ‘of the hairs continue somewhat obliquely through it.

d.Exposure 2 min. The ‘roots ‘can now be seen passing right through the exocuticle and disappearing in the outer layers of the endocuticle. Magnification of a-d × 2850.

(N.B. The arrow marks the same spot on the section in all four photographs.)

PLATE 7

Fig. e. Section through abdominal spiracular rosette of adult Aphelocheirus. Method as above. Aperture 5 mm., exposure 3 min. Magnification × 2850.

Fig. f. Transverse section of sternal cuticle of fifth instar nymph of Aphelocheirus, method as above. Aperture 5 mm., exposure 2 min. Note absence of hair pile, simple epicuticle, dark exocuticle showing transverse line suggesting pore canals towards inner margin, and thick unpigmented endocuticle in which a large number of lamellae are faintly visible. Magnification × 2850.

Fig. g. Electron microscope photograph of gelatin-formvar preparation of plastron surface of abdominal sternum of adult Aphelocheirus. The numerous small somewhat triangular bodies represent the tips of the plastron hairs. Magnification × 91,200.

Fig. h. Electron microscope ‘silhouette ‘of a small part of a section of abdominal sternum showing tip of a single hair projecting from the edge of the hair pile which is itself quite opaque to the electron beam. Magnification × 120,000.